ML RSI - imaclone.x

Ultimate Multi-Resolution RSI
Developed by imaclone.x.
Last Updated: August 21st 2025

A single indicator that fuses my ML-RSI.ai pipeline with a classic multi-timeframe RSI. One script, dual-resolution oscillators if desired, plus a machine-learning similarity engine and modular signal-processing layers.

What it does

* Primary RSI augmented with KNN similarity engine (K, lookback, weighting). Feature embeddings include RSI magnitude, RSI momentum, volatility surface, regression slope, and price momentum vectors.
* Adaptive smoothing stack: Kalman filter recursion, Double EMA cascades, or ALMA convolution.
* Multi-resolution control for the primary oscillator timeframe.
* Optional *second* RSI projected from any timeframe for hierarchical confluence.
* Advanced visuals: upper/lower thresholds, midline, background regime highlighting, crossovers, and B/S event labels.
* Color architectures: None, Trend-Following (50-line bifurcation), or Impulse (band-breach). Optional bar tinting for full-chart context.

Inputs (groups)

* Timeframe Settings: primary + secondary RSI TF/lengths.
* Levels & Visuals: thresholds, highlights, cross events, B/S markers.
* RSI Base: smoothing toggle, MA class, ALMA sigma.
* KNN Machine Learning: enable, K neighbors, historical window, feature dimensionality, ML weighting.
* Advanced Filtering: method + intensity.
* Coloring: None, Trend-Following, Impulse.

Signals

* B flag when ML-RSI crosses upward through the lower threshold.
* S flag when ML-RSI crosses downward through the upper threshold.
* Secondary RSI = higher-timeframe confirmation, not standalone trigger.

Usage notes

* Raise ML weight + feature dimensionality for deeper similarity recognition; lower them for classic oscillator behavior.
* Kalman recursion delivers adaptive, low-lag smoothing; Double EMA and ALMA yield stronger dampening.
* Typical config: intraday primary RSI + higher-TF secondary RSI for regime anchoring.

Changelog

* v6 merge: Unified CM-style MTF RSI framework with my KNN-enhanced kernel and filter stack. One composite indicator replaces multiple scripts.

Credits

* MTF band logic inspired by earlier open-source frameworks.
* ML kernel and implementation by imaclone.x.

Disclaimer
For research and algorithmic experimentation only. No signals guaranteed.
And please kindly, for the love of God, DYOFR.